System Info
optimum==2.1.0
optimum-onnx==0.1.0
Python 3.12.7
MacOS 26.3.1
Who can help?
@echarlaix @IlyasMoutawwakil I'm tagging you because I ran into this issue while trying to convert to ONNX format
Information
Tasks
Reproduction (minimal, reproducible, runnable)
Link to my dataset https://storage.googleapis.com/models-backend-artifacts/models/furnished-classifier.zip
But should reproduce with any other timm image classification fine tuned models.
To get an error enough to use script from official documentation:
!optimum-cli export onnx --model furnished-classifier --task image-classification onnx_furnished_classifier
Error:
Traceback (most recent call last):
File "/Users/antonkharkhonov/projects/ss/ai/propvision/custom-models/runpod-app/.venv/bin/optimum-cli", line 8, in <module>
sys.exit(main())
^^^^^^
File "/Users/antonkharkhonov/projects/ss/ai/propvision/custom-models/runpod-app/.venv/lib/python3.12/site-packages/optimum/commands/optimum_cli.py", line 219, in main
service.run()
File "/Users/antonkharkhonov/projects/ss/ai/propvision/custom-models/runpod-app/.venv/lib/python3.12/site-packages/optimum/commands/export/onnx.py", line 264, in run
main_export(
File "/Users/antonkharkhonov/projects/ss/ai/propvision/custom-models/runpod-app/.venv/lib/python3.12/site-packages/optimum/exporters/onnx/__main__.py", line 325, in main_export
model = TasksManager.get_model_from_task(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/antonkharkhonov/projects/ss/ai/propvision/custom-models/runpod-app/.venv/lib/python3.12/site-packages/optimum/exporters/tasks.py", line 1175, in get_model_from_task
model = model_class(f"hf_hub:{model_name_or_path}", pretrained=True, exportable=True)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/antonkharkhonov/projects/ss/ai/propvision/custom-models/runpod-app/.venv/lib/python3.12/site-packages/timm/models/_factory.py", line 114, in create_model
pretrained_cfg, model_name, model_args = load_model_config_from_hf(
...
Repository Not Found for url: https://huggingface.co/furnished-classifier/resolve/main/config.json.
Please make sure you specified the correct `repo_id` and `repo_type`.
If you are trying to access a private or gated repo, make sure you are authenticated. For more details, see https://huggingface.co/docs/huggingface_hub/authentication
Invalid username or password.
From what I found manually in optimum code, only remote model loading is allowed intentionally for timm models:
https://github.com/huggingface/optimum/blob/main/optimum/exporters/tasks.py#L1177
Expected behavior
Successful conversion of the model to ONNX format
System Info
Who can help?
@echarlaix @IlyasMoutawwakil I'm tagging you because I ran into this issue while trying to convert to ONNX format
Information
Tasks
examplesfolder (such as GLUE/SQuAD, ...)Reproduction (minimal, reproducible, runnable)
Link to my dataset https://storage.googleapis.com/models-backend-artifacts/models/furnished-classifier.zip
But should reproduce with any other timm image classification fine tuned models.
To get an error enough to use script from official documentation:
Error:
From what I found manually in optimum code, only remote model loading is allowed intentionally for timm models:
https://github.com/huggingface/optimum/blob/main/optimum/exporters/tasks.py#L1177
Expected behavior
Successful conversion of the model to ONNX format